# Paired t-test. H-E-L-P.

#### Malcolm

##### New Member
Hi!

I'm really, really hoping someone can help me with this, because I've been trying to figure it out on my own, and I'm getting NOWHERE.

Here goes: I'm a student working on a paper regarding the use of cholesterol- lowering meds and cholesterol-leves in blood. My paper is basically based upon several published articles. I'm doing no research of my own (I'm a freshman pharmacy- student), and therefore the data I have, is restricted to the data given in the articles. Anyway, what I'm doing, is that I am trying to calculate if the reduction of blood- cholesterol is statistically significant after starting to use cholesterol-lowering meds, in different groups of patients (age, medical history, gender, etc). I have the mean pre- and posttreatment values of blood-cholesterol of these patient, and the mean pre- and post standard deviation of this. I figured I had to use a paired t-test to calculate p. The problem is, though, that I only have the mean values of the groups, I do not have the values of each patient, so that I cannot calculate "s" in the formula for the paired t-test. The more I read about this, the more confused I get, so I would be eternally grateful if someone could point me in the right direction. How do i calculate p from these mean- values?
I hope I'm making sense, English is not my native language.
Thank you!

#### duskstar

##### New Member
Are you trying to calculate a p-value for each study or an overall one (in some form of meta-analysis)? Do you know n (the number of observations) for each study?

#### Malcolm

##### New Member
Hello.

I'm trying to calculate p for each study. I know the number of observations for each study, but I don't have every patients individual value. I only have the mean pre- and postoperative value, and the standard deviation for these.

#### Dason

##### Ambassador to the humans
Well... you can't do a paired t-test then. Don't know what else to tell you.